fix(0.21.10-followup): actually drop --full-depth (BSD sed silent no-op in #63)#64
Conversation
…kills restructure Companion to broomva/skills PR #11 (merge ab2d05b) which removed the root SKILL.md. The CLI's default search now descends into skills/<name>/ without the flag. Reverted: skills-roster.md (26 cmds) + skill-graduate.sh (4 generated cmds). VERSION 0.21.9 → 0.21.10. 20/20 tests still pass.
…ate CLI v0.21.10 (PR #63 merge eefd463) shipped with a CHANGELOG entry claiming 'dropped --full-depth from 26 + 4 commands' but the actual file changes were silently lost: BSD sed (macOS) doesn't support \b word-boundary in substitutions, so 's| --full-depth\b||g' was a no-op. Only VERSION + CHANGELOG landed. This PR closes the gap: portable sed 's| --full-depth||g' applied; counts verified at 0/0 in both files; tests still 20/20. After this lands, v0.21.10 matches its CHANGELOG claim.
|
Warning Review limit reached
More reviews will be available in 57 minutes and 53 seconds. Learn how PR review limits work. Your organization has run out of usage credits. Purchase more in the billing tab. ⌛ How to resolve this issue?After more reviews become available, a review can be triggered using the We recommend that you space out your commits to avoid hitting the rate limit. 🚦 How do rate limits work?CodeRabbit enforces hourly rate limits for each developer per organization. Our paid plans include higher PR review limits than trial, open-source, and free plans. In all cases, reviews become available again over time. During sustained high-volume PR review activity, CodeRabbit may temporarily slow when the next review becomes available. Please see our Fair Usage Limits Policy for further information. ℹ️ Review info⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (4)
✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
…clude check Pre-existing on main (from #64) — the "Lint shell + JSON templates" gate inherited red on this PR. $ex is a glob pattern from EXTRA_EXCLUDES; quoting the RHS would break the pattern match the exclude check depends on, so the correct fix is a disable directive, not a quote. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
… verdict (#65) * feat(0.22.0): wire RCS control loop on bootstrap + doctor §23 closure verdict Close the split-brain: onboard.sh wired the control loop via install-rcs-stability.sh; bootstrap.sh (the /bstack bootstrap command) never did, leaving freshly-bootstrapped workspaces with governance files but an OPEN loop (no L0/L1 audit hooks, no .control/audit/, no closure-arc definitions). doctor reported all of that as soft [info], so the open loop was invisible. Dep-Chain (P14): - upstream: scripts/install-rcs-stability.sh (reused, unchanged) → install-l3-stability.sh (writes to $WORKSPACE, verified no PWD bug); assets/templates/arcs.yaml.template; compute-budget-status.sh / compute-arc-status.sh. - downstream: scripts/onboard.sh (already wired — now converges with bootstrap); scripts/repair.sh (already offers install-rcs); tests/canary/01; every workspace that runs `bstack bootstrap` or `bstack doctor`. Changes: - bootstrap.sh Phase 3.5: call install-rcs-stability.sh (BSTACK_SKIP_RCS=1 escape; `|| true` preserves non-blocking contract under set -e). Phase 2: scaffold .control/arcs.yaml from template. - doctor.sh §23: control-loop closure verdict (substrate-absent / wired-but-idle / wired+running+closing) from W/R/C signals. Soft by default; BSTACK_LOOP_STRICT=1 promotes wired-but-idle to a hard --strict gap. §19 already hard-gates a wired loop that is genuinely diverging, so §23 does not double-count. - canary/01: assert PostToolUse hook + .control/audit/ + .control/arcs.yaml + L0/L1 markers (14/14). references/new-workspace-flow.md + CHANGELOG + VERSION 0.21.10→0.22.0. Verified: canary 01-05 green; arcs-validation/repair-merge-hooks/onboard suites green; §23 renders correctly across all three states + strict path. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix(lint): SC2053 disable on intentional glob in skill-graduate.sh exclude check Pre-existing on main (from #64) — the "Lint shell + JSON templates" gate inherited red on this PR. $ex is a glob pattern from EXTRA_EXCLUDES; quoting the RHS would break the pattern match the exclude check depends on, so the correct fix is a disable directive, not a quote. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix(§23): W-check reads settings.local.json too (dogfood-surfaced) Dogfooding 0.22.0 on the stimulus repo surfaced a false-negative: stimulus's tracked .claude/settings.json uses repo-relative vendored hooks, so the L0/L1 audit hooks (which carry an absolute bstack path) belong in the gitignored settings.local.json — which Claude Code merges at runtime. §23's W-check only grepped settings.json, so it reported "NOT wired" while the loop was actually running. Now checks both files. Verified: §23 on stimulus flips to "wired + running + closing" with hooks in settings.local.json; canary 14/14. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Carlos D. Escobar-Valbuena <devteam@getstimulus.ai> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
v0.21.10 (PR #63) shipped with a CHANGELOG claiming '
--full-depthdropped from 26 + 4 commands' but the actual file edits were silently lost — BSD sed (macOS) doesn't support\bword-boundary in substitutions, so the regex was a no-op. Only VERSION + CHANGELOG landed.This PR closes the gap with a portable
sed 's| --full-depth||g'(verified 0/0 in both files; 20/20 tests pass). After merge, v0.21.10 matches its CHANGELOG.